I vibe coded a $20K/month mobile app in 14 days

Download ~20 apps in the niche plus beautifully designed apps for inspiration. Screenshot every page of every onboarding and lay them in one line in a Figma file to see every question and graph; pick favorites from each and redesign them in your own theme. Spend as long on onboarding as on core features, since ~90% of users only see onboarding if the app is paywalled. Onboarding should: evoke emotion (purchases are emotional), show the strongest benefits and make buying simple, feel personalized, and include charts/graphs so it feels scientific and proven. Payout's onboarding asks for activity and notifications, shows social proof, then the paywall.

Paywall at the end of onboarding shows a weekly and a yearly option, designed to prioritize the yearly plan because it has the highest lifetime value per customer. Uses RevenueCat for subscription analytics and pricing tests.

Before coding, write a text document explaining the app's data plus example JSON describing each attribute, and give it to the AI (Claude/Cursor) so it doesn't guess. Build the core functionality first, skipping onboarding. Drop design screenshots into Claude so it writes the full screen; iterate when it gets details wrong but don't be too picky: ship a version fast rather than a perfect app. Now uses only Claude (Claude Code) for coding.

Partnered with a content creator already in Payout's niche to promote the app to his existing audience, relying on the trust followers already have in him. Aim for relatable, viral content that feels natural and unprocessed.

Once a creator video performs well, get more videos in the same format from other creators. The UGC builds a library of ads whose likely performance is known from organic results. Goal is always to reach paid ads, the most consistent revenue driver. On Facebook, entertaining videos get cheaper delivery, but the ad must still push the product, so balance entertainment with conversion.

Picks app ideas that tap core desires (be healthier, make more money, look attractive). Payout finds class action settlements users qualify for, so paying for it feels like an investment that will return money.

After his first app (social, Hotspot Events) took 6-8 months and failed, concluded social apps are very hard to scale and moved to utilities and tools, which are easier for one person to scale. Advice: build something simple with one to three good features and a great onboarding.

Modified versions of existing apps are already validated; just make them your own. For brand-new ideas, search TikTok and Instagram to see whether people talk about the problem and ask 'how do I solve this?' in comments, which also signals an easy marketing path. Because building is now fast, skip the waitlist step and build the app.